Looking for information on transistor based designs

[PRR]

> is there any point in learning how transistors work?

You mean orbitals and holes and bandgaps? NO! I'm only starting to learn that stuff, and I'm not sure I believe it.

Transistors DO work by magic. Or at least you can design many-many nifty things without any INternal theory. While some of these books 4-page-gloss the physics, you can skip that.

EXternally you need to know that a happy transistor will show about 0.6V Vbe, that collector voltage is whatever the external load forces it to, that colector and emitter current are essentially equal, that base current is much-much (Beta) less. AND that Beta is WILDLY uncertain. (In Germanium you must also be aware of Ico....) That's the kind of books we are discussing. How to do something useful with a wacky magical device.

[cthulhudarren]

For me personally I'd like to learn more about the circuit analysis. The DC is usually pretty easy, but AC is not. Also, it always seems that the spec sheets don't have the information like Beta at the operating points I'm looking for. Actually, how to read, interpret and use a transistor spec sheet is not easy! With tubes the circuit gain is set by the the circuit around the tubes, and with transistors it's the properties of the transistor at the DC operating point in the circuit. So it's not easy for me to understand what gain the designer was wanting to get out of a transistor gain stage.

[PRR]

> With tubes the circuit gain is set by the the circuit around the tubes, and with transistors it's the properties of the transistor at the DC operating point in the circuit.

I think you have that backward.

The maximum gain of a tube triode is not large, 15 to 100 depending which tube. In almost all work, we strive to get as much of that as possible, 15 to 50. So knowing the tube gives a very close guess at the gain.

The maximum voltage gain of a BJT transistor is incredible, over 1,000. So much that we can't handle it. OTOH the current gain is modest, 100-200, and VERY variable, leading to low and uncertain input impedance. In almost all work, we strive to NOT depend on transistor properties, limiting the available voltage-gain with resistors, and padding-up the low impedance.

The most simple technique is a large emitter resistor. Now the voltage gain is nearly the ratio of load resistance to emitter resistance.

Because we want a low DC gain (for consistent bias with varying devices) and usually some higher AC gain, this leads to a R-C-R emitter network with a large capacitor. There is another less-stable technique with nearly-grounded emitter and "inverter type" negative feedback. Again the AC gain is nearly the ratio of two resistors.